Lunch & learn: So…what is Artificial Intelligence (AI)?
In our daily Lunch & Learn sessions you will hear from experts as they provide a brief overview of topics impacting the global financial industry. Today’s topic is all about artificial intelligence, or AI. Lunch will be provided so you can eat while you learn.
The advent of digital technologies like automation and computerised systems has changed our environment substantially by providing enormous amounts of data. How can AI help us to make the most of this new digital era? Apart from the public perception of science fiction-like applications of AI such as robots, AI has a lot of practical applications with the aim of reducing costs, increasing benefits, optimising resource usages and improving the accuracy of decisions. These aims can be achieved in a range of business functions from marketing, sales and supply chain management to product development, human resources and public services.
After a brief general introduction to AI, demystifying some common concepts, e.g. explaining differences between ‘automation’ and ‘intelligence’, Professor Aickelin, Head of the School of Computing and Information Systems at the University of Melbourne, will outline three of the most common AI methods in Fintech - evolutionary computation, neural networks and clustering.
Evolutionary computation gets inspiration from natural evolution and adaptation processes where a random set of feasible solutions is generated. This initial set iteratively updates by removing less desired solutions and producing new solutions stochastically. The iterative process continues to find the best value for the objective function.
Neural networks are AI systems inspired by the way neurons interact in the brain. They need high computer processing power and large sets of data to analyse different input data such as image, video and speech. Modern neural networks are referred to as ‘deep learning’.
Clustering is an unsupervised AI technique to group observations in different subsets based on their characteristics. In clustering, grouping occurs in a way that items among each group are similar while the groups are distant.
Professor Aickelin’s talk will conclude with a look at some recent relevant legislation around data and privacy in Europe and what consequences this may have on the use of AI. For instance, what does it mean if algorithms have to be ‘explainable’ in the future?